HEREDITY, AND THE ORIGIN OF SPECIES. 29 
characters, the change is a reversionary one, and the qual- 
ities that appear are really latent in the entire plant, and 
only need some stimulus to awaken them, or some agency 
to weaken the dominancy of the prevalent characters. 
In the case of the appearance of characters not hitherto 
borne by the main stock, the case is not so clear, especially 
as we feel fairly certain that the saltations do ensue, in seed- 
mutants at least, in single cells. Here the theoretical side 
of the case seemed least supported by facts and I set about 
supplying the deficiency with what success you may pres- 
ently judge. 
Bud-sports are usually found on the basal portions of 
main stems of shoots, or in that pat of the main axis, which 
often shows juvenile structures, so that we may say that 
these breaks in heredity are very frequent in the closing 
part of the embryonic period, although sometimes occur- 
ring much later. On the other hand, the changes which 
result in the formation of seed-mutants must ensue in the 
very closing stages of the sporophyte, and here is encoun- 
tered a feature not recognized in bud-mutations, that of a 
fairly constant frequency, which in itself would suggest 
internal causes. It was determined however to attempt 
to induce or influence mutation in one of these periods. « It 
was deemed impracticable to attempt this in the early 
life of the sporophyte and attention was given entirely to 
the stage immediately preceding the reducing divisions. 
Omitting the detail of technique I may say that strong 
osmotic reagents, and weak solutions of stimulating min- 
eral salts were injected into ovaries in such manner that 
unfertilized ovules were subjected to the action of the 
fluids, which killed many of them, but which gave the 
much desired results in a few. In Lamarckiana and bien- 
nis the frequency and character of the known mutants was 
unaffected, but in the progeny of the latter species was 
found a single individual constituting a type hitherto un- 
